Filters have long been used in the processing of electrical signals. For example, in communications applications, such as microwave applications, it is desirable to filter out the smallest possible passband and thereby enable dividing a fixed frequency spectrum into the largest possible number of bands.
Such filters are of particular importance in the telecommunications field (microwave band). As more users desire to use the microwave band, the use of narrow-band filters will increase the actual number of users able to fit in a fixed spectrum. Of most particular importance is the frequency range from approximately 800-2,200 MHz. In the United States, the 800-900 MHz range is used for analog cellular communications. Personal communication services are planned for the 1,800 to 2,200 MHz range.
Historically, filters have fallen into three broad categories. First, lumped element filters have used separately fabricated air wound inductors and parallel plate capacitors, wired together to form a filter circuit. These conventional components are relatively small compared to the wave length, and accordingly, make for a fairly compact filter. However, the use of separate elements has proved to be difficult to manufacture, resulting in large circuit to circuit variations. The second conventional filter structure utilizes three-dimensional distributed element components. These physical elements are sizeable compared to the wavelength. Coupled bars or rods are used to form transmission line networks which are arranged as a filter circuit. Ordinarily, the length of the bars or rods is 1/4 or 1/2 of the wavelength at the center frequency of the filter. Accordingly, the bars or rods can become quite sizeable, often being several inches long, resulting in filters over a foot in length. Third, printed distributed element filters have been used. Generally, they comprise a single layer of metal traces printed on an insulating substrate, with a ground plane on the back of the substrate. The traces are arranged as transmission line networks to make a filter. Again, the size of these filters can become quite large. These filters also suffer from various responses at multiples of the center frequency.
Historically, filters have been fabricated using normal, that is, non-superconducting materials. These materials have inherent lossiness, and as a result, the circuits formed from them have varying degrees of loss. For resonant circuits, the loss is particularly critical. The Q of a device is a measure of its power dissipation or lossiness. Resonant circuits fabricated from normal metals in a microstrip or stripline configuration have Qs at best on the order of four hundred. See, e.g., F. J. Winters, et al., "High Dielectric Constant Strip Line Band Pass Filters", IEEE Transactions On Microwave Theory and Techniques, Vol. 39, No. 12, December 1991, pp. 2182-87.
With the discovery of high temperature superconductivity in 1986, attempts have been made to fabricate electrical devices from high-temperature-superconductor materials. The microwave properties of the high temperature superconductors (HTSCs) has improved substantially since their discovery. Epitaxial superconductive thin films are now routinely formed and commercially available. See, e.g., R. B. Hammond et al, "Epitaxial Tl.sub.2 Ca.sub.1 Ba.sub.2 Cu.sub.2 O.sub.8 Thin Films With Low 9.6 GHz Surface Resistance at High Power and Above 77.degree. K", Applied Physics Letters, Vol. 57, pp 825-27 (1990). Various filter structures and resonators have been formed from HTSCs. Other discrete circuits for filters in the microwave region have been described. See, e.g., S. H. Talisa, et al., "Low- and High-Temperature Superconducting Microwave filters," IEEE Transactions on Microwave Theory and Techniques, Vol. 39, No. 9, September 1991, pp. 1448-1554.
Devices with zero resistance should have an infinite Q. However, even superconductive devices are not perfectly lossless at high frequencies. However, they do have exceedingly high Qs. For example, a thallium superconductor strip line resonator at 8.45 GHz has been measured with a Q of 26,000 as compared to a Q of literally a few hundred for the best conventional metal resonator. See, e.g., F. J. Winters, et al., "High Dielectric Constant Strip Line Band Pass Filters" cited above.
Vairous filter structures have been formed utilizing significant superconductive components. See, e.g., U.S. Pat. No. 5,616,538 "STAGGERED RESONATOR ARRAY". In many applications keeping filter structures to a minimum size is very important. This is particularly true of high-temperature superconductor (HTS) filters where the available size of usable substrates is generally limited. In the case of narrow-band microstrip filters (e.g., bandwidths of the order of 2 percent, but more-especially 1 percent or less) this size problem can become quite severe. In narrow-band microstrip filters substantial differences between even- and odd- mode wave velocities exist when the substrate dielectric constant is large. This can create relatively large forward coupling between the resonators thereby presenting a need for large spacings between the resonators in order to obtain the required narrow bandwidth. See, G. L. Matthaei and G. L. Hey-Shipton, "Concerning the Use of High-Temperature Superconductivity in Planar Microwave Filters," IEEE Trans. on MTT, vol. 42, pp. 1287-1293, July 1994. This may make the overall filter structure unattractively large or, perhaps, impractical or impossible for some situations.
FIG. 1 shows a two-resonator comb-line filter 10 realized in a stripline configuration so the even- and odd-mode velocities on the coupled lines will be equal (thus, preventing forward coupling). The two resonators 11 are grounded at the sidewall 12, and in this example the input and output couplings 13 are provided by tapped-line connections. This structure would have no passband at all if it were not for the "loading" capacitors Cr 14. From the equivalent circuit for a comb-line filter it can be seen why this happens. See, G. L. Matthaei, L. Young, and E. M. T. Jones, Microwave Filters, Impedance-Matching Networks, and Coupling Structures, Artech House Books, Dedham, Mass., 1980, pp. 497-506 and 516-518.
Since the resonators are shorted at one end, when loading capacitors are zero (Cr=0) the resonators are resonant when they are a quarter-wavelength long. As seen from their open-circuited ends, they look like shunt-connected, parallel-type resonators which would yield a passband at this frequency. However, there is also an odd-mode resonance in the region between the lines which acts like a bandstop resonator connected in series between two shunt resonators. This creates a pole of attenuation at the same frequency that a passband would otherwise occur. Thus, the potential passband is totally blocked. However, if loading capacitors, Cr&gt;0, are added at the ends of the resonators, the resonator lines are shortened in order to maintain the-same passband frequency. This shortens the length of the slot between the lines and causes the pole of attenuation to move up in frequency away from the passband.
In general, the more capacitive loading used, the further the pole of attenuation would be above the passband, and the wider the passband of the filter can be. If only small loading capacitors Cr are used, a very narrow passband can be achieved even though the resonators are physically quite close together. Similar operation also occurs if more resonators are present. If the structure in FIG. 1 is realized in a microstrip configuration, the performance is considerably altered because of the different even- and odd-mode velocities, though some of the same properties exist in modified form.
FIG. 2A shows a common form of hairpin-resonator bandpass filter 20. See, E. G. Cristal and S. Frankel, "Hairpin-Line and Hybrid Hairpin-Line/Half-Wave Parallel-Coupled-Line Filters," IEEE Trans. MTT, vol.
MTT-20, pp. 719-728, November 1972. The filter 20 can be thought of as an alternative version of the parallelcoupled-resonator filter first introduced by S. B. Cohn in "Parallel-Coupled Transmission-Line-Resonator Filters," IRE Trans. PGMTT, vol. MTT-6, pp. 223-231 (April 1958), except that here the resonators 21 are folded back on themselves. See G. L. Matthaei, L. Young, and E. M. T. Jones, Microwave Filters, Impedance-Matching Networks, and Coupling Structures, Artech House Books, Dedham, Mass., 1980, pp. 472-477). Note that in FIG. 2A the orientations of the hairpin-resonators 21 alternate (i.e. neighboring resonators face opposite directions). This results in quite strong coupling which makes this structure capable of considerable bandwidth. However, in the case of narrow-band filters, particularly for microstrip filters on a high-dielectric substrate, this structure is undesirable as it may require quite large spacings between the resonators 21 to achieve a desired narrow bandwidth.
FIG. 2B shows another common form of hairpin-resonator filter 22. See, M. Sagawa, K. Takahashi, and M. Makimoto, "Miniaturized Hairpin Resonator Filters and Their Application to Receiver Front-End MIC's," IEEE Trans. MTT, vol. 37, pp. 1991-1997 (December 1989). In this case the open-circuited ends 23 of the resonators 24 are considerably foreshortened and a strongly capacitive gap 25 is added to bring the remaining structure into resonance. The resonators are then semilumped, the lower part 26 being inductive and the upper part 27 being capacitive. The coupling between resonators 24 is almost entirely inductive, and it makes little difference whether adjacent resonators are inverted with respect to each other or not. Hence, as is shown in FIG. 2B, these resonators are usually made to have the same orientation (i.e. neighboring resonators face the same direction). If the resonators have sufficiently large capacitive loading these resonator structures can be quite small, but, typically, their Q is inferior to that of a full hairpin resonator. Also, there will normally be no resonance effect in the region between the resonators so that the coupling mechanism cannot be used to generate poles of attenuation beside the passband in order to enhance the stopband attenuation.
Therefore, the need for compact, reliable, and efficient narrow-band filters with very high Q resonator which can be manufactured with consistency remains unsatisfied. Despite the clear desirability of improved electrical circuits, including the known desirability of converting circuitry to include superconducting elements, room remains for improvement in devising alternate structures for filters. It has proved to be especially difficult to substitute high temperature superconducting materials in conventional circuits to form superconducting circuits without severely degrading the intrinsic Q of the superconducting films. Among the problems encountered are radiative losses and tuning, which remain despite the clear desirability of improved filters. As is described above, size has remained a concern, especially for narrow-band filters. Also, power limitations arise in certain structures. Despite the clear desirability for forming microwave filters for narrow-band applications, to permit efficient use of the frequency spectrum, a need remains for improved designs capable of achieving those results in an efficient and cost effective manner.